Latest / The Steve Harvey Morning Show / Using AI to Rethink the Way Work Gets Done

Transcript

Using AI to Rethink the Way Work Gets Done · The Steve Harvey Morning Show

• 0:02
  Speaker 1: Hello everyone. This is Smart Talks with IBM, a podcast
• 0:07
  Speaker 1: from Pushkin Industries, iHeartMedia and IBM about what it means
• 0:12
  Speaker 1: to look at today's most challenging problems in a new way.
• 0:16
  Speaker 1: I'm Malcolm Glapwell. Today I'm chatting with Rob Thomas, the
• 0:25
  Speaker 1: senior vice president of IBM Cloud and Data, where his
• 0:29
  Speaker 1: responsibility is bringing new ideas to life. But despite being
• 0:36
  Speaker 1: on the cutting edge of these technologies, he still has
• 0:39
  Speaker 1: an appreciation for age old problems.
• 0:46
  Speaker 2: There's a rabbit and a beaver and they're staring at
• 0:49
  Speaker 2: the Hoover Dam, and the beaver says to the rabbit, no,
• 0:53
  Speaker 2: I didn't build it, but it's based on an idea
• 0:55
  Speaker 2: of mind. And the point of that story is there's
• 0:59
  Speaker 2: ideas are a dime that doesn't great. Everybody's got a
• 1:05
  Speaker 2: bunch of ideas.
• 1:06
  Speaker 1: By the way, we're too quick to dismiss the beaver.
• 1:10
  Speaker 1: He's right, but have you seen beaver dams?
• 1:12
  Speaker 2: I mean, I know he is right. It was his idea,
• 1:16
  Speaker 2: but he had nothing to do with the giant cement
• 1:18
  Speaker 2: Hoover new.
• 1:22
  Speaker 1: In my interview with Rob, we'll touch on the importance
• 1:25
  Speaker 1: of the cloud during the pandemic and how IBM has
• 1:28
  Speaker 1: been playing a part in vaccine distribution Stay tuned. I,
• 1:38
  Speaker 1: for one, had no idea about what it means to
• 1:41
  Speaker 1: be the senior vice president of IBM Cloud and Data,
• 1:43
  Speaker 1: so I asked Rob to break it down for me
• 1:45
  Speaker 1: in layman's terms.
• 1:47
  Speaker 2: We built software, and software is the lingua franca of
• 1:53
  Speaker 2: our time. Anything that will get done in businesses and
• 1:58
  Speaker 2: even interaction with consumers is going to be done with software.
• 2:03
  Speaker 2: It's really the language of everything that's happening in the world.
• 2:07
  Speaker 2: That's what we build. We are focused on doing that
• 2:10
  Speaker 2: for businesses.
• 2:11
  Speaker 1: So how long have you been at.
• 2:13
  Speaker 2: IBM twenty years or twenty one, I guess to be precise.
• 2:17
  Speaker 2: And I started in consulting, and then I moved into
• 2:22
  Speaker 2: our semiconductor business, and I was doing consulting. And the
• 2:27
  Speaker 2: moment that really changed my whole career was doing work
• 2:31
  Speaker 2: with Nintendo where we were designing the microprocessor for the
• 2:36
  Speaker 2: Nintendo Wii, and I realized, we're going to do this
• 2:40
  Speaker 2: one time, but then they're going to be building software
• 2:43
  Speaker 2: that will get copied billions of times and used by
• 2:46
  Speaker 2: people all over the world. Maybe I'm not in the
• 2:49
  Speaker 2: right business. And that really peaked my curiosity around software,
• 2:54
  Speaker 2: which then led me to move into the IBM software business,
• 2:58
  Speaker 2: where I've been for most of my career at this point.
• 3:01
  Speaker 2: So I've been in software total twelve thirteen years.
• 3:06
  Speaker 1: But you have seen I'm guessing, so twelve years in software.
• 3:09
  Speaker 1: Am I right in thinking that more has happened in
• 3:12
  Speaker 1: those twelve years of software then happened in the entire
• 3:15
  Speaker 1: history of software before that? Is that a fair statement? Close?
• 3:19
  Speaker 2: It's I'd say close. It's certainly the rate and pace
• 3:22
  Speaker 2: of innovation has increased now as actually something hasn't been
• 3:26
  Speaker 2: an outcome. Maybe that's a different question. Yeah, but if
• 3:29
  Speaker 2: you think about you know, software dates way back to
• 3:33
  Speaker 2: even the first mainframe that we ever built in the fifties,
• 3:37
  Speaker 2: So a lot of good things have been happening in
• 3:39
  Speaker 2: software for a long time, but the rate and pace
• 3:42
  Speaker 2: is a level that we've never seen, and that's certainly
• 3:45
  Speaker 2: been what is accelerated in the last decade.
• 3:47
  Speaker 1: Yeah, I mean, I remember my dad was a mathematician
• 3:52
  Speaker 1: the Naresity of Waterloo. I remember coming home as a kid,
• 3:55
  Speaker 1: going into his office and seeing stacks of computer cards.
• 3:58
  Speaker 1: So in my lifetime I have gone from looking at
• 4:02
  Speaker 1: stacks of computer cards to something far more so. I mean,
• 4:07
  Speaker 1: I am aware of just how fast this fast, this
• 4:11
  Speaker 1: this pace is gone.
• 4:13
  Speaker 2: And it'll be different a year from now. Right, that's
• 4:16
  Speaker 2: how fast this is moving.
• 4:18
  Speaker 1: Let's zero in on that a little bit. What's shifting
• 4:23
  Speaker 1: right now? Imagine I'm a client. I come to you,
• 4:26
  Speaker 1: and I say, you know, I want to be prepared
• 4:28
  Speaker 1: for next year and a year after next. What should
• 4:31
  Speaker 1: be at the top of my mind.
• 4:35
  Speaker 2: Let me give you a quick story, if you don't mind.
• 4:37
  Speaker 2: There was a time in the US where you could
• 4:40
  Speaker 2: not easily get from one city to another. And at
• 4:45
  Speaker 2: that point, back in the nineteen fifties, there was a
• 4:48
  Speaker 2: decision that said, let's actually build the infrastructure to connect
• 4:53
  Speaker 2: every city in America. And the result was fifty plus
• 4:58
  Speaker 2: years of work, four hundred million dollar and we now
• 5:00
  Speaker 2: have forty eight thousand miles of highways that connects all
• 5:04
  Speaker 2: these cities. But the real impact is more profound than that,
• 5:09
  Speaker 2: because you were able to eliminate traffic at intersections by
• 5:13
  Speaker 2: building overpasses. There are all these second order businesses that
• 5:18
  Speaker 2: were built. Hotels, gas stations, the salty snacks that you
• 5:24
  Speaker 2: buy in a gas station, fast food rest areas, so
• 5:28
  Speaker 2: an entire economy was built around the idea that the
• 5:33
  Speaker 2: first step was just to connect all the cities in
• 5:36
  Speaker 2: the US. And that's what's happening right now with software.
• 5:41
  Speaker 2: It is connecting businesses and individuals in a way that
• 5:45
  Speaker 2: we've never been connected before, and we are just at
• 5:48
  Speaker 2: the beginning of all the second order effects that will
• 5:52
  Speaker 2: come as a result of them. And the biggest problem
• 5:55
  Speaker 2: in software it's data. Just like you had all these
• 5:59
  Speaker 2: disparate cities and you were building highways to connect those cities,
• 6:05
  Speaker 2: every company has all these different data sets all over
• 6:09
  Speaker 2: the place and it's a really hard problem. But AI
• 6:13
  Speaker 2: is not going to be a reality for businesses until
• 6:16
  Speaker 2: the data problem is solved. That's one thing that I
• 6:21
  Speaker 2: spend a lot of time on right now.
• 6:25
  Speaker 1: Dig in from with that into that meeting of that phrase,
• 6:28
  Speaker 1: the data problem.
• 6:29
  Speaker 2: I think every individual wants any company they interact with,
• 6:34
  Speaker 2: whether it's their local bank or restaurant, or the local cleaners,
• 6:40
  Speaker 2: whatever it may be, they want that business to know them.
• 6:45
  Speaker 2: It's the whole idea of when you had towns where
• 6:48
  Speaker 2: there was just one general store and the owner knew you,
• 6:51
  Speaker 2: they knew what you wanted. I think everybody wants that
• 6:54
  Speaker 2: level of engagement, and that is what software enables. And
• 7:01
  Speaker 2: the basis of that is data. And the biggest problem
• 7:05
  Speaker 2: every business faces today is how do I understand my data?
• 7:10
  Speaker 2: What it tells me about my customers, what it tells
• 7:12
  Speaker 2: me about my products. So this is fundamentally about how
• 7:16
  Speaker 2: do we live in a better way.
• 7:19
  Speaker 1: You're talking about that. I'm a big company and I
• 7:22
  Speaker 1: have different sets of data, and they're all in different places,
• 7:26
  Speaker 1: and they don't speak to each other, and I can't
• 7:28
  Speaker 1: combine them and make sense. And is that what you
• 7:30
  Speaker 1: mean by the data problem?
• 7:32
  Speaker 2: Correct? And even if I can combine them and connect them,
• 7:37
  Speaker 2: the data is not in a usable form. You know,
• 7:41
  Speaker 2: one data, says m Gladwell, the other one says Malcolm
• 7:44
  Speaker 2: g Is that the same person? Maybe? Maybe not. It's
• 7:48
  Speaker 2: really hard because these systems have been built up over time.
• 7:52
  Speaker 2: We do work with a company called Wonderman Thompson. Story
• 7:56
  Speaker 2: that they shared with me just this month was doing
• 8:01
  Speaker 2: work with Peloton. So Peloton collects a lot of data
• 8:05
  Speaker 2: what you call first party data from a bike or
• 8:08
  Speaker 2: the tread. I think you're a runner if I recall
• 8:11
  Speaker 2: and WPP Wonderman Thompson has all this third party data,
• 8:16
  Speaker 2: which is what do they know about consumers? So just
• 8:19
  Speaker 2: to connect those two data sets, build predictive models, and
• 8:24
  Speaker 2: then to turn that into an advertising campaign. The AI
• 8:27
  Speaker 2: part is actually relatively easy. It's actually connecting the data,
• 8:31
  Speaker 2: rationalizing the data, cleaning the data. That's the really hard
• 8:35
  Speaker 2: part that nobody talks about because all we ever see is,
• 8:39
  Speaker 2: you know, the outcome.
• 8:40
  Speaker 1: Yeah, well, just so I understand this, it's super interesting.
• 8:43
  Speaker 1: So let's imagine you Rob are a Peloton user, and
• 8:46
  Speaker 1: so we have a data stream that comes from the
• 8:48
  Speaker 1: bike which says that you bike. Let's just say for them,
• 8:53
  Speaker 1: I'm going to flatter you an hour and a half
• 8:55
  Speaker 1: a day at some insane pace.
• 8:58
  Speaker 2: And neither of which are true. But keep going. I
• 9:00
  Speaker 2: did do a half hour today, but it was a
• 9:02
  Speaker 2: very slow pace, I gotta tell you so.
• 9:04
  Speaker 1: And I'm looking at your whatever it is I'm collecting.
• 9:08
  Speaker 1: I'm assuming Peloton collects a lot of sort of physiological
• 9:11
  Speaker 1: and use data from the bike, and from that we
• 9:14
  Speaker 1: can generate a rough sense of who you are, how
• 9:18
  Speaker 1: what your athletic interests are, how fit you are, all
• 9:21
  Speaker 1: those kinds of things. And Wonderman's Old Shop wants to know,
• 9:26
  Speaker 1: how can I use that picture of the kind of
• 9:28
  Speaker 1: athlete you are to help bring you the kinds of
• 9:32
  Speaker 1: ad messages that you'll respond to. Is that a fair?
• 9:35
  Speaker 1: Is that the problem?
• 9:36
  Speaker 2: It could be bringing it to me? But it's more
• 9:38
  Speaker 2: likely because obviously they de anonymize all this data. It's
• 9:41
  Speaker 2: more of all, right, so, how do we find somebody
• 9:43
  Speaker 2: else that's like rob? What are the attributes of that person?
• 9:46
  Speaker 2: And then how do we relate to them in a
• 9:49
  Speaker 2: way that makes it feel like we're talking to them
• 9:52
  Speaker 2: as opposed to talking to a cohort or a group.
• 9:57
  Speaker 2: The number one prediction that most companies want to make
• 10:00
  Speaker 2: is what's going to happen to my sales next month
• 10:03
  Speaker 2: or the month after or the month after. And what
• 10:07
  Speaker 2: we found is that tends to be a product of
• 10:10
  Speaker 2: as many as fifty or one hundred different inputs. How
• 10:12
  Speaker 2: many people are visiting the website, how many people are
• 10:15
  Speaker 2: calling the call center, how many sales calls if I
• 10:18
  Speaker 2: have a face to face salesforce are they making? How
• 10:21
  Speaker 2: many marketing campaigns am I running? If you take all
• 10:25
  Speaker 2: of these different data points, which is often in fifty
• 10:28
  Speaker 2: or one hundred, you feed those into a model, then
• 10:34
  Speaker 2: the first month, you see how close was the model.
• 10:36
  Speaker 2: Then you adjust second month, you see how close was
• 10:39
  Speaker 2: the model? And these models get really good over time.
• 10:43
  Speaker 2: And we think we can help companies predict their financial
• 10:47
  Speaker 2: performance in a month and a quarter in a year
• 10:50
  Speaker 2: based on all these different data sources, all these different inputs.
• 10:53
  Speaker 2: That's pretty valuable. Tot's say every company.
• 10:56
  Speaker 1: Yeah, so IBM, what's IBM's role in that you've described
• 11:00
  Speaker 1: that problem to me? What does you? Guys come in
• 11:02
  Speaker 1: and you say, will do what?
• 11:06
  Speaker 2: A couple of years ago I started. I was trying
• 11:08
  Speaker 2: to think about what is the right metaphor so that
• 11:11
  Speaker 2: I can educate our customers on this and built this
• 11:17
  Speaker 2: concept that I call the AI ladder. To think of
• 11:25
  Speaker 2: it as steps that you take up a ladder towards AI.
• 11:29
  Speaker 2: The bottom rung is collect data. So you have to
• 11:32
  Speaker 2: be able to collect all your data. I'll use a
• 11:35
  Speaker 2: library analogy. This is just you have to get books.
• 11:38
  Speaker 2: You have to get books into the library that's collecting.
• 11:41
  Speaker 2: Next is you have to organize that data and the
• 11:45
  Speaker 2: a lot back to library analogy. That's the card catalog.
• 11:48
  Speaker 2: So where are all the different data sets? I might
• 11:51
  Speaker 2: have five copies of the same data, How do I
• 11:53
  Speaker 2: know that's the same copy, maybe ones checked out, maybe
• 11:56
  Speaker 2: ones on microfilm. These are actually all problems that exist
• 11:59
  Speaker 2: in businesses. So you've got to collect data, you've got
• 12:02
  Speaker 2: to organize data, then you have to analyze the data.
• 12:05
  Speaker 2: So you're actually starting to do data science machine learning
• 12:10
  Speaker 2: in the library metaphor that's where you're displaying your bestseller
• 12:13
  Speaker 2: list or you're displaying popular magazine titles. And then the
• 12:18
  Speaker 2: top of the ladder is what I call infuse. So
• 12:21
  Speaker 2: then how do you take those models and infuse them
• 12:26
  Speaker 2: into a business process. So it's those four steps the ladder.
• 12:30
  Speaker 2: You have to collect, organize, analyze, infuse. We build software
• 12:36
  Speaker 2: that helps customers with each of the rungs of that ladder,
• 12:39
  Speaker 2: helps them do the collection. We actually build what we
• 12:42
  Speaker 2: call a data catalog to help you organize your data.
• 12:46
  Speaker 2: So we help them with all rungs of that ladder.
• 12:48
  Speaker 2: Because ultimately, then you've probably heard of IBM Watson, that
• 12:52
  Speaker 2: is our AI platform. Once you've done those things, you
• 12:57
  Speaker 2: can use AI and get really great outcomes.
• 13:04
  Speaker 1: Imagine if someone from the White House came to you
• 13:09
  Speaker 1: and said, we're about to do something we've never done
• 13:12
  Speaker 1: in this or haven't done in this country for seventy years,
• 13:16
  Speaker 1: which is try and vaccinate everybody in the shortest possible time.
• 13:20
  Speaker 1: We have multiple sets of three and eventually probably four
• 13:24
  Speaker 1: or five different kinds of vaccines being administered by tens
• 13:28
  Speaker 1: of thousands of local municipalities to people who have a
• 13:31
  Speaker 1: wide ranging set of risk factors, urgency, pre existing conditions,
• 13:38
  Speaker 1: going on, you know, on and on and on and on.
• 13:42
  Speaker 1: Can you help us do this as efficiently and cost
• 13:47
  Speaker 1: effectively and socially consciously as possible? Is that a kind
• 13:53
  Speaker 1: of task that you're talking.
• 13:54
  Speaker 2: About now that is in part as much a logistics
• 13:59
  Speaker 2: problem as it is a data problem. Let me describe
• 14:02
  Speaker 2: to you one of the data problems though that exists
• 14:04
  Speaker 2: around this because we're doing the work with CVS on
• 14:09
  Speaker 2: the COVID vaccine rollout. Oh yes, and so if you're
• 14:15
  Speaker 2: CVS where you're actually administering, their biggest problem is everybody
• 14:20
  Speaker 2: has a question. CVS can't hire enough people to answer
• 14:24
  Speaker 2: the ten questions you have, the ten questions I have,
• 14:27
  Speaker 2: the twenty questions your cousin has. They came to us
• 14:32
  Speaker 2: and said, can we use AI to respond to all
• 14:36
  Speaker 2: the inquiries we're getting and actually help route people to
• 14:40
  Speaker 2: where they can figure out they can get the vaccine
• 14:42
  Speaker 2: when they're eligible. So we built an AI agent for
• 14:46
  Speaker 2: them that is now dealing with the vaccine rollout every day.
• 14:52
  Speaker 2: That starts with data. They have a place that they
• 14:55
  Speaker 2: store data about different questions. We've got models that we
• 15:00
  Speaker 2: have trained on language, meaning we can understand different types
• 15:04
  Speaker 2: of questions. What's really inferred versus implied versus what is stated.
• 15:09
  Speaker 2: That's a real data problem. That's where we've spent majority
• 15:14
  Speaker 2: of our time looking at this current situation.
• 15:17
  Speaker 1: So you would that when you say it's a data problem,
• 15:20
  Speaker 1: meaning that you started by trying to anticipate, by looking
• 15:26
  Speaker 1: at the data and using that to try and anticipate
• 15:28
  Speaker 1: all the possible questions that someone might ask, Is that
• 15:30
  Speaker 1: what you're correct?
• 15:32
  Speaker 2: Yes, and then training a machine learning model based on
• 15:36
  Speaker 2: those inputs so that when the question was asked, we
• 15:40
  Speaker 2: had a high probability of giving the right answer. Yeah.
• 15:43
  Speaker 1: How long did it take you to build that system?
• 15:47
  Speaker 2: Now, this is the wonders of modern software. To your
• 15:51
  Speaker 2: question on acceleration, we did this in forty five days.
• 15:54
  Speaker 1: Are you serious? Yeah, it's insane. How and how many
• 15:58
  Speaker 1: people worked on it?
• 16:02
  Speaker 2: Twenty five thirty somewhere in that room. It's not a
• 16:04
  Speaker 2: huge group.
• 16:06
  Speaker 1: Wow.
• 16:08
  Speaker 2: Their thing about systems like this is you hope it's
• 16:11
  Speaker 2: really good on day one, but you know for sure
• 16:14
  Speaker 2: it's going to be better on day ten, it's going
• 16:16
  Speaker 2: to be better again on day twenty. These are learning systems.
• 16:20
  Speaker 2: They do get better over time. And the thing is
• 16:23
  Speaker 2: with the really difficult problems. And this is this is
• 16:27
  Speaker 2: why I like to talk about AI as giving human superpowers.
• 16:30
  Speaker 2: Most people want to say it replaces humans. I actually
• 16:32
  Speaker 2: think gives them superpowers because in these cases you start
• 16:36
  Speaker 2: to move the harder problems to the humans, and so
• 16:39
  Speaker 2: therefore your your customer satisfaction goes up because people are
• 16:43
  Speaker 2: getting their problems resolved. Would you ever get to one
• 16:45
  Speaker 2: hundred percent? Probably not, because there's always going to be
• 16:47
  Speaker 2: something that's too difficult for the AI to handle. But
• 16:50
  Speaker 2: I think you can keep moving it up for sure.
• 16:52
  Speaker 1: Yeah. Or maybe given what you've just said, would it
• 16:55
  Speaker 1: be more fair to say you don't want to get
• 16:57
  Speaker 1: to one hundred that you want to resis serve a
• 17:01
  Speaker 1: certain category of problem for a human to human interaction,
• 17:04
  Speaker 1: because that might be ultimately more satisfying to the question.
• 17:10
  Speaker 2: We have that discussion a lot, and certainly in the
• 17:12
  Speaker 2: ones that I've worked on, that's that's typically the case,
• 17:15
  Speaker 2: because let's not forget these are businesses, and the goal
• 17:19
  Speaker 2: of most businesses is to sell something. So sometimes the
• 17:22
  Speaker 2: best way to sell something is to really help somebody
• 17:24
  Speaker 2: with their problem and then show them how your other
• 17:27
  Speaker 2: product can make their life even easier.
• 17:30
  Speaker 1: When you think back in the cases that you have
• 17:33
  Speaker 1: kind of problems that that your group has been asked
• 17:36
  Speaker 1: to solve at IBM of last couple of years, what
• 17:39
  Speaker 1: was the hardest.
• 17:42
  Speaker 2: I don't know that I could name a single thing
• 17:44
  Speaker 2: that's harder than others. The ones that are the most
• 17:47
  Speaker 2: time consuming, things like regulatory compliance. If you're a bank,
• 17:52
  Speaker 2: you've got a lot of different regulations that you have
• 17:55
  Speaker 2: to live up to. And it's easy to help build
• 18:00
  Speaker 2: AI that can make loan decisions yes or no, good idea,
• 18:05
  Speaker 2: bad idea, eliminate bias from that decision. That's very doable.
• 18:10
  Speaker 2: Am I compliant with the regulations of where that individual
• 18:14
  Speaker 2: is based? Because they're in a zip code, or they're
• 18:16
  Speaker 2: in a state, they're in a country. Those problems get
• 18:18
  Speaker 2: really difficult because you're kind of connecting you know, reams
• 18:23
  Speaker 2: of legality to a day to day business process. Get
• 18:28
  Speaker 2: those get really difficult.
• 18:30
  Speaker 1: Has any customer ever come to you with a problem
• 18:33
  Speaker 1: that you guys said, we can't solve that.
• 18:38
  Speaker 2: We are way too curious to ever give up that easily. Yeah,
• 18:43
  Speaker 2: it's more of you know, it's the it's the cheap, fast,
• 18:46
  Speaker 2: and good triangle. If you've heard that, you know you
• 18:49
  Speaker 2: only get two of those. Do you want it cheap
• 18:51
  Speaker 2: and fast, it's probably not good. If you want it
• 18:53
  Speaker 2: good and fast, it's probably not cheap. If you want
• 18:56
  Speaker 2: it cheap and good, it's probably not fast. I think
• 18:59
  Speaker 2: all of these situations come down to that triangle.
• 19:04
  Speaker 1: So you have a group of people who work on
• 19:05
  Speaker 1: these kinds of problems, and I'm curious, what do you
• 19:08
  Speaker 1: look for when you're bringing someone on too that team.
• 19:14
  Speaker 1: Is there a set of skills associated with dealing with
• 19:18
  Speaker 1: this area of the application of AI to these very
• 19:21
  Speaker 1: complicated data fields. Is there a specific set of skills
• 19:26
  Speaker 1: that are crucial and rare hard find?
• 19:30
  Speaker 2: The skill that's easy to test for is do you
• 19:36
  Speaker 2: have the technical ability? Do you understand Python? Do you
• 19:41
  Speaker 2: understand machine learning? You can kind of see from somebody's
• 19:46
  Speaker 2: body of work and what they've studied. Do they have
• 19:48
  Speaker 2: that skill? Part where it gets harder is the empathy question.
• 19:55
  Speaker 2: Can you actually understand a situation, understand a user, and
• 20:00
  Speaker 2: empathize with what they're trying to do such that you're
• 20:03
  Speaker 2: not just building a model for a robot, You're actually
• 20:05
  Speaker 2: building this for a human on some end, that one's hard,
• 20:09
  Speaker 2: harder to test for. And then the third one is
• 20:12
  Speaker 2: I would just call it curiosity. How widely read is somebody?
• 20:17
  Speaker 2: Do they understand business business problems? Because those kind of
• 20:22
  Speaker 2: softer skills, those make a huge difference when you're solving
• 20:27
  Speaker 2: these kinds of problems. So it's easy to test for
• 20:29
  Speaker 2: the first. The other two are a little harder to
• 20:32
  Speaker 2: test for, and the best data scientists in the world
• 20:36
  Speaker 2: have all three of those.
• 20:37
  Speaker 1: Let's talk about the cloud. I see this word hybrid cloud.
• 20:46
  Speaker 1: I don't know what it means. So can you start
• 20:49
  Speaker 1: by telling me what it means and then fit this
• 20:51
  Speaker 1: into the conversation we've been having.
• 20:55
  Speaker 2: So any company that's been around for more than three years,
• 21:00
  Speaker 2: maybe five, they've got somewhere that they keep their data
• 21:07
  Speaker 2: and they keep the different technology that they have, and
• 21:11
  Speaker 2: in many cases that's in their office, or that's in
• 21:13
  Speaker 2: a data center right right near their office. They've also
• 21:18
  Speaker 2: started over time to start to build new data sets
• 21:23
  Speaker 2: or new software in a public cloud, something from you,
• 21:29
  Speaker 2: something inside of IBM Cloud or Amazon Web Services or
• 21:35
  Speaker 2: Microsoft Azure. The minute that you have more than one environment,
• 21:42
  Speaker 2: you have a hybrid cloud, whether whether you know it
• 21:44
  Speaker 2: or not. So think of it as I've got multiple
• 21:48
  Speaker 2: data sets in multiple places to kind of back to
• 21:50
  Speaker 2: the US Highway example, or I've got software applications in
• 21:55
  Speaker 2: multiple places. You have to get that to act like
• 22:00
  Speaker 2: a single technology environment. That is the essence of hybrid cloud,
• 22:05
  Speaker 2: which is I can manage that as a single environment.
• 22:08
  Speaker 2: The average company now has five different environments cloud wise,
• 22:12
  Speaker 2: it acts like one. I can connect the data the.
• 22:15
  Speaker 1: Average company has. Is that by design because they feel
• 22:20
  Speaker 1: it's safe for Is that just because the hodgepodge nature
• 22:24
  Speaker 1: in which we grow our it needs means that we
• 22:27
  Speaker 1: end up being all over the place.
• 22:31
  Speaker 2: It's because there's a lot of people that work in
• 22:34
  Speaker 2: every company, and everybody wants their own thing. That's how
• 22:37
  Speaker 2: it happens. So this department started in their own data center,
• 22:42
  Speaker 2: This department started on IBM Cloud. This department wanted a
• 22:47
  Speaker 2: CRM system from Salesforce. This department wanted to use Azure.
• 22:51
  Speaker 2: It's human nature. People just go do what they want
• 22:55
  Speaker 2: to do. And you wake up one day and you realize, hey,
• 22:58
  Speaker 2: we've got a lot of different cloud in midroms. And
• 23:01
  Speaker 2: so if you're storing your customer data with Salesforce and
• 23:04
  Speaker 2: you've got these three other environments, how do you get
• 23:06
  Speaker 2: the customer data to inform you know what you're doing
• 23:10
  Speaker 2: in the other parts of your business. That's a hybrid
• 23:12
  Speaker 2: cloud problem.
• 23:13
  Speaker 1: Yeah, and how how hard of a problem is that?
• 23:19
  Speaker 1: I mean, as a total naive outsider, I would have said, oh,
• 23:23
  Speaker 1: surely all these cloud businesses would have made it really
• 23:26
  Speaker 1: easy to share stuff in one place with stuff you've
• 23:28
  Speaker 1: got in other place. Is that not true?
• 23:31
  Speaker 2: Unfortunately the opposite is true, because for the pure play
• 23:35
  Speaker 2: public cloud providers, the incentive was actually the opposite. It's
• 23:39
  Speaker 2: Hotel California. For them. You can bring your stuff in,
• 23:41
  Speaker 2: but you know you can, you can check in, but
• 23:43
  Speaker 2: you will never let you check out, and they charge
• 23:46
  Speaker 2: actually enormous fees if you want to get your data out.
• 23:50
  Speaker 2: So it's a bit of a strategy tax for them
• 23:52
  Speaker 2: to make it easy it's also a hard problem just
• 23:55
  Speaker 2: because you're trying to connect different data sets. Do you
• 23:59
  Speaker 2: have one card catalog that connects all these different sources.
• 24:04
  Speaker 2: It's actually not easy to do. And what happens when
• 24:07
  Speaker 2: you don't do that then you end up rebuilding everything
• 24:09
  Speaker 2: and so suddenly you're storing all the same data five times.
• 24:12
  Speaker 2: That gets very expensive.
• 24:14
  Speaker 1: So let's imagine we're having this conversation in twenty twenty six.
• 24:20
  Speaker 1: Give me your sense of where will be would what
• 24:23
  Speaker 1: would we be talking about? Five years from now.
• 24:26
  Speaker 2: We'll probably be having very similar discussions as possible. Technology
• 24:30
  Speaker 2: will be more advanced, but a lot of the problems
• 24:33
  Speaker 2: we've talked about, let's be honest, these have been around
• 24:35
  Speaker 2: for quite a while. There's a story this guy, Charles Towns,
• 24:41
  Speaker 2: he was the inventor of the laser, and he tells
• 24:43
  Speaker 2: his story. There's a rabbit and a beaver and they're
• 24:46
  Speaker 2: staring at the Hoover dam and the beaver says to
• 24:49
  Speaker 2: the rabbit, no, I didn't build it. But it's based
• 24:52
  Speaker 2: on an idea of mind. And the point of that
• 24:55
  Speaker 2: story is there's you know, ideas are a diamond doesn't Yeah,
• 24:59
  Speaker 2: so great? Everybody's got a bunch of ideas.
• 25:04
  Speaker 1: By the way, we're too quick to dismiss the beaver.
• 25:08
  Speaker 1: He's right, but and have you seen Beaver Dams.
• 25:10
  Speaker 2: I mean, I know he is right. It was his idea,
• 25:14
  Speaker 2: but he had nothing to do with the giant cement
• 25:16
  Speaker 2: Hoover dam. The reason I'd share that story is a
• 25:24
  Speaker 2: lot of people have ideas now about what they can do,
• 25:27
  Speaker 2: But what's going to make a difference five years from
• 25:29
  Speaker 2: now is what do you go try and do. And
• 25:34
  Speaker 2: I encourage companies that you've got to be willing to
• 25:37
  Speaker 2: have a pretty high failure rate, knowing that if you
• 25:41
  Speaker 2: go try a bunch of things, you know, maybe only
• 25:44
  Speaker 2: half of them will work out. I mean, if I
• 25:46
  Speaker 2: look at AI today, so there's five major things I
• 25:49
  Speaker 2: see companies doing generally successfully. It's customer service. We talked
• 25:53
  Speaker 2: about that. It's financial budgeting, It's regulatory compliance. We talked
• 25:59
  Speaker 2: about that. That one's harder. It's employee experience hiring that
• 26:04
  Speaker 2: type of thing. And it's using AI to run their
• 26:07
  Speaker 2: IT systems. So using software to run the systems. Those
• 26:11
  Speaker 2: are the five big things today. I actually think those
• 26:16
  Speaker 2: five things will still be the topic in twenty twenty six,
• 26:19
  Speaker 2: but will be a lot more advanced on each of
• 26:21
  Speaker 2: those because today it's a little bit we're doing it
• 26:24
  Speaker 2: for the first time, whereas we'll be a much more
• 26:26
  Speaker 2: advanced as we get out to twenty twenty six. I
• 26:29
  Speaker 2: do think quantum computing will be commercialized at that point.
• 26:32
  Speaker 2: That's pretty revolutionary. So more to come on that one.
• 26:37
  Speaker 1: Let's end on some more case studies. Tell me a
• 26:41
  Speaker 1: couple of examples of people you've worked with where the
• 26:44
  Speaker 1: outcome is was really exciting or unexpected.
• 26:49
  Speaker 2: Or we've worked with Sprint T Mobile. They have this
• 26:56
  Speaker 2: classic problem of they've got to do after market service
• 27:01
  Speaker 2: for all the different telecom equipment that they sell and
• 27:06
  Speaker 2: the data that they have on those different systems, the
• 27:11
  Speaker 2: warranty when they were built, how they're running, it's spread
• 27:14
  Speaker 2: across a thousand different different data sources. We were able
• 27:20
  Speaker 2: to build an AI system for them that sits across
• 27:24
  Speaker 2: those systems, that was able to intelligently route how they
• 27:28
  Speaker 2: do all of their after market service. So do you
• 27:32
  Speaker 2: and I feel that in our day to day life, Well,
• 27:36
  Speaker 2: we feel that if they don't fix things, then it's
• 27:39
  Speaker 2: obvious because there's an outage or something that doesn't work.
• 27:41
  Speaker 2: But it was something that they had so much data
• 27:44
  Speaker 2: on this, they could have never done this by i'd
• 27:47
  Speaker 2: say a typical approach. So these are the kinds of
• 27:51
  Speaker 2: things that the average consumer doesn't see every day, but
• 27:55
  Speaker 2: they do make a difference in our life.
• 27:58
  Speaker 1: And you're talking about things like what like cell towers.
• 28:02
  Speaker 2: Or yeah, it could be cell towers, or it could
• 28:04
  Speaker 2: be you know, the power cable, not the power cable,
• 28:07
  Speaker 2: the power box is sitting next to the cell tower.
• 28:09
  Speaker 2: It could be any of those things.
• 28:10
  Speaker 1: Oh, I see. So you're so they have all of
• 28:12
  Speaker 1: these systems that might have been bought at different times,
• 28:15
  Speaker 1: made by different people, installed by different people, and so
• 28:19
  Speaker 1: what you want to do is to give them a
• 28:21
  Speaker 1: system that allows them to look at them all in
• 28:23
  Speaker 1: real time and figure out where there might be an issue.
• 28:27
  Speaker 2: Yes, we call it predictive maintenance. Right, Okay, all the
• 28:31
  Speaker 2: signs are that there's about to be a problem on
• 28:33
  Speaker 2: this one, they go out there, they check it out,
• 28:35
  Speaker 2: yep lo and behold, there is a problem.
• 28:37
  Speaker 1: Yeah, I have two cars, can you build one for me?
• 28:40
  Speaker 2: So we haven't scaled down to that level quite yet,
• 28:45
  Speaker 2: but stay tuned. We're open to it.
• 28:46
  Speaker 1: Why not, Why are you neglecting I'm the ultimate end user.
• 28:50
  Speaker 1: I'm one guy with two cars.
• 28:54
  Speaker 2: That's a good question. We'll bring it to you soon.
• 28:58
  Speaker 1: We have any have any professional sports teams work with you?
• 29:03
  Speaker 2: Guys Toronto Raptors publicity on that a few years ago.
• 29:09
  Speaker 1: You're You're I'm Canadian. This is that's my team. You're
• 29:12
  Speaker 1: warming my heart right now of course. Well this has
• 29:15
  Speaker 1: been really fun. Thank you so much.
• 29:19
  Speaker 2: Yeah, I'm Malcolm, appreciate it.
• 29:24
  Speaker 1: I'd love to help out the Toronto Raptors if I
• 29:26
  Speaker 1: had the chance. Thanks again to Rob Thomas for an
• 29:30
  Speaker 1: intriguing conversation about data and the cloud. Smart Talks with
• 29:38
  Speaker 1: IBM is produced by Emily Rosstek with Carle Migliore, edited
• 29:43
  Speaker 1: by Karen Shakerji engineering by Martin Gonzalez, mixed and mastered
• 29:49
  Speaker 1: by Jason Gambrel. Music by Gramoscope. Special thanks to Malli Soshia,
• 29:56
  Speaker 1: Andy Kelly, Mio Label, Jacob Weisberg, Heather Fane, Eric Sander,
• 30:00
  Speaker 1: Maggie Taylor, and the teams at eight Bar and IBM.
• 30:05
  Speaker 1: Smart Talks with IBM is a production of Pushkin Industries
• 30:09
  Speaker 1: and iHeartMedia. You can find more Pushkin podcasts on the
• 30:14
  Speaker 1: iHeartRadio app Apple podcasts or wherever you like to listen.
• 30:19
  Speaker 1: I'm Malcolm Gladwell, See you next time.